Additions/improvements in existing code for gesture recognition

I need some additions in existing code. I have a C++ code in xcode (it’s also an object in max/msp, implemented with the max/msp API) and a max/msp interface that does gesture recognition by using machine learning algorithms. The system is able to be learned by a reference gesture, and then to be recognized by the same gesture made from the same or another person. Also, what is already implemented is that in order the system to recognize the gesture, the user can put a threshold (constant value), in which if the gesture is inside this value, the system recognizes the gesture as correct. What I need is the following addition, not to have only the constant value for all the gesture data, but the system to be able to read in real time a table with different float values for each input data per time. With real-time I mean that the system will read these values at the same time that the gesture is being executed, and will check if the gesture is inside these values for all input data, by recognizing it as correct. I’ll provide you with recorded data and text files containing float values, but you can test it also with real data if you want, by yourself if you have a camera, like kinect or leap motion. For more information, there is a more detailed description after you bid. Also, the code and the max/msp interface should come with an explanatory and detailed documentation, explaining step by step the implementation, so that I can understand exactly how it works! Also, I’m planning to make more improvements in this code in the near future. Thank you in advance!!